A Novel Approach to Measurement Drift Phenomenon in Terrain Aided Navigation

• Published in: 2011 Fourth International Conference on Intelligent Computation Technology and Automation Vol. 1; pp. 559 - 563
• Main Authors: Shidan Li, Liguo Sun, Xin Li, Desheng Wang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.03.2011
• Subjects: Atmospheric measurements; Atmospheric modeling; Computational modeling; Mathematical model; Measurement Drift Model; Particle Filter; Particle filters; Particle measurements; Recursive Bayesian Filter; Sea measurements; Terrain Aided Navigation
• ISBN: 1612842895; 9781612842899
• DOI: 10.1109/ICICTA.2011.151

Terrain Aided Navigation (TAN) system has been widely used in aerial and underwater vehicles for its autonomy, anti-electromagnetic interference (anti-EMI) and all-weather advantages. TAN system is also a multi-sensor system usually consists of radar altimeter, barometric altimeter, laser scanner, echo sounder, inertial navigation system (INS) and so on. It fuses the information from different sensors to highly improve the positioning accuracy that can not be acquired by a single positioning system, such as INS. However, the positioning accuracy of TAN can be easily affected by the altimeter measurement errors which may caused by atmosphere turbulence or water current. This paper focuses on the effort to overcome the influences of these aspects. We propose a high dimensional measurement drift system model that incorporates measurement bias and drift aspects, which can model the influences above. The recursive Bayesian filter is used for analytic solutions, while a modified SIR particle filter is in charge of numerically calculating the results. The paper also examines the traditional model without measurement drift for comparison. From the simulation results, the positioning accuracy of the measurement drift model has improved by approximate one order of magnitude than the traditional model.